Predictive Audiences: Taking Personalization to the Next Level

Image showing personalization.

by Liron Goren Snai

posted on 05-20-2020

In these uncertain times, optimization and personalization can be powerful tools. Across Adobe Experience Cloud, we’ve seen brands embrace their personalization strategies due to shifting customer needs and concerns. What remains stagnant is the challenge around personalization at scale – delivering tailored, meaningful and relevant messages that customers actually want. But delivering customized experiences widely is often a painstaking process for organizations, as customers are interacting with content across an increasing number of channels.

Leveraging artificial intelligence and machine learning technology, Adobe Audience Manager is releasing new ways for brands to deliver personalization at scale, seamlessly, while honoring user choice and control. With Predictive Audiences, brands can maximize the impact of marketing initiatives by classifying an unknown audience, i.e. people that are visiting a brand’s site but not yet categorized into a segment, into distinct personas, in real-time. Marketers can classify users into audiences to personalize across channels and devices, and improve intelligence and personalization for unknown audiences and users who have limited trait associations.

Design, classify, differentiate

Through tailored customization of each model based on specific needs and use cases, Predictive Audiences solves for a number of challenges brands are facing today. After allowing marketers to define specific categories or personas by which they want to group their audience, machine learning will match an unclassified user’s propensities against an existing segment and predict which persona this user should most likely belong to based on the information the user shares. This classification happens in real-time, coupled with a distinct prediction, to personalize across channels and devices leveraging Adobe Audience Manager’s identity management capabilities.

For example, a marketer can leverage Predictive Audiences to classify website visitors into different categories to personalize on-site in real-time, and ultimately improve the reach of their campaigns. Separately, an advertiser can classify unknown audiences by behavioral attributes to retarget customers on ad platforms with personalized messages to increase conversion rates, or a brand can separate customers based on where they are in their journey (discovery, engage, purchase, retain) to offer personalized offers that are relevant.

Adobe integrates privacy-by-design as a fundamental principle and provides various controls to enterprise customers to honor the data privacy choices of their customers. Predictive Audiences adopts Data Export Controls, a patented Data Governance feature in Adobe Audience Manager. With Data Export Controls, marketers can create predictive segments with complete control and auto enforcement on the type of data that is collected and exported for these segments based on business requirements, data usage agreements or organizational data governance. The model also classifies only those users who are opted in, based on their preferences, as with any other segment in Audience Manager.

Early feedback

Multiple brands have been testing Predictive Audiences in beta, including Sprint.

Sprint is involved in a powerful digital transformation centered around personalization. Kevin Day, Martech Manager at Sprint, notes the benefits for their team, _“_Adobe Audience Manager allows Sprint to better understand the needs of customers when they visit our website. The journey of each customer is very clear allowing us to move quickly and provide a personalized experience.”

Predictive Audiences is available for all Adobe Audience Manager customers to use today.

Topics: Personalization, News

Products: Audience Manager